Subject: Re: kern/33185: kva shortage problems
To: None <kern-bug-people@netbsd.org, gnats-admin@netbsd.org,>
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
List: netbsd-bugs
Date: 05/11/2006 13:24:11
The following reply was made to PR kern/33185; it has been noted by GNATS.
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
To: yamt@mwd.biglobe.ne.jp
Cc: gnats-bugs@NetBSD.org, tech-kern@NetBSD.org,
tsutsui@ceres.dti.ne.jp
Subject: Re: kern/33185: kva shortage problems
Date: Thu, 11 May 2006 20:08:44 +0900
yamt@mwd.biglobe.ne.jp wrote:
> > > a concern is kva fragmentation.
> >
> > I'm not a VM expert, but "kva fragmentation" reminds me
> > STE shortage on 68040 and sun3x pmaps...
> i'm not sure what's "STE shortage". can you explain with MI terms? :-)
Well, sun3x uses 7-6-6-13 bits 3-level MMU, and memory
for each Segment Table Entry is allocated statically.
Recently (for a year?) I often got "get_x_tabe: out of X tables."
panics on my 3/80 and I just thought it was caused by kva
fragmentation.
(yes, just sun3x pmap should be fixed, but it's not easy...)
> i don't think my patch makes kva fragmentation much worse.
Ok, then there is no problem on your patch.
Thank you for your explanation.
---
Izumi Tsutsui